Forecast Of Highway Freight Index Based On Network Attention And Baidu Index

The highway freight index is a measure of the fluctuation of the road transport market,and it has an important predictive function to the road transport industry of our country China’s road freight index monitoring system is not complete yet,and it is difficult for government management departments to monitor the road freight rate fluctuations accurately.But the maturity of big data and cloud computing technology in the Internet era makes it feasible to use network attention to predict highway freight index.At present,network attention has become a hot issue in the field of behavior management research,and network attention has become a compass for predicting economic fluctuations.Behavioral finance believes that individual behavior affects market price changes.In the road transportation market,production and management personnel pay attention to specific road transportation market-related factors.These concerns are projected on the Internet and transform into a search record of a certain keyword.The attention behavior represented by search records affects the price changes in the road transportation market.The research uses Baidu engine searching records(Baidu index)as a proxy variable of the network attention of the road transportation market,supplemented by emerging big data technology,and proposes a accurate prediction method for highway freight index.To use the Baidu index to predict the highway freight index accurately,the paper must answer the relationship between the two factors first.Therefore,the ER attribute entity principle of the database and the traditional measurement principle of highway freight index is used as a reference.The random forest and Pearson correlation coefficients are used to screen out the most suitable Baidu index keywords to form the input sequence for prediction;secondly,ADF test and cointegration test are used to verify the stable relationship between Baidu index and highway freight index,which provides logic for using Baidu index to predict highway freight index.Then,the paper optimized the window of the searched Baidu index to eliminate the influence of extreme values between the data through rolling time domain optimization,and predicted the changes in the highway freight market more accurately;finally,different models are tested,and the prediction accuracy is evaluated and the best extreme learning machine prediction method is got.Research shows that the prediction of highway freight index based on Baidu index improved the accuracy of the prediction model and can predict changes in the highway transportation price market in advance and accurately by using extreme learning machine.After selecting the RBF kernel function,the optimal accuracy of the extreme learning machine is reached,which has better predictive ability than the linear kernel function.Under window rolling time domain optimization,the volatility of the highway freight index was predicted in advance.In the process of constructing the framework of the prediction model,the network attention represented by the Baidu index is correlated with the road freight index,and it is found that the Baidu index of semi-trailer price,road bridge fee,used truck,oil0 and truck loan are more effective in predicting the road freight index.The research introduces the perspective of network attention into the field of road freight rate prediction,which has certain value for the highway transport market managers to supervise the changes of road freight rate and macro-control the road transport market.